<div dir="ltr">hummm...<div>OK, thanks for information!</div><div>Very good.</div><div><br></div><div>Good luck!</div></div><div class="gmail_extra"><br><br><div class="gmail_quote">2013/8/13 Tatsuo Ishii <span dir="ltr"><<a href="mailto:ishii@postgresql.org" target="_blank">ishii@postgresql.org</a>></span><br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div class="im">> No problem...<br>
> If your active server is the number 2 then run the command (killall -9<br>
> pgpool) himself.<br>
> You will see that all processes die and the interface with the ip delegate<br>
> will remain active. This is the case...<br>
<br>
</div>That is only applicatable to pgpool-II 3.2.<br>
<br>
3.3's watchdog monitors parent pgpool process is alive. If it's gone,<br>
release the ip and tells standby watchdog that it goes to "down"<br>
status.<br>
<div class="HOEnZb"><div class="h5">--<br>
Tatsuo Ishii<br>
SRA OSS, Inc. Japan<br>
English: <a href="http://www.sraoss.co.jp/index_en.php" target="_blank">http://www.sraoss.co.jp/index_en.php</a><br>
Japanese: <a href="http://www.sraoss.co.jp" target="_blank">http://www.sraoss.co.jp</a><br>
<br>
> 2013/8/13 Tatsuo Ishii <<a href="mailto:ishii@postgresql.org">ishii@postgresql.org</a>><br>
><br>
>> > You said: "I killed server2 pgpool-II parent process by kill -9"<br>
>> > Try "killall -9 pgpool" in active server...<br>
>><br>
>> Sorry for confusion but in my case server2 is the active.<br>
>> --<br>
>> Tatsuo Ishii<br>
>> SRA OSS, Inc. Japan<br>
>> English: <a href="http://www.sraoss.co.jp/index_en.php" target="_blank">http://www.sraoss.co.jp/index_en.php</a><br>
>> Japanese: <a href="http://www.sraoss.co.jp" target="_blank">http://www.sraoss.co.jp</a><br>
>><br>
>> > The stand by will grow up the delegate IP and the same will still on old<br>
>> > active server and will conflict.<br>
>> > So I am using a shell script in crontab to check this and restart pgpool<br>
>> if<br>
>> > necessary.<br>
>> ><br>
>> ><br>
>> > 2013/8/12 Tatsuo Ishii <<a href="mailto:ishii@postgresql.org">ishii@postgresql.org</a>><br>
>> ><br>
>> >> > On 08/05/13 14:07, Fernando Buzon wrote:<br>
>> >> >><br>
>> >> >> FINAL:<br>
>> >> >> Like I said, all is working nice.<br>
>> >> >> And now I am with the 2 pgpools up and working again.<br>
>> >> >> The escaled pgpool is pgpool-01.<br>
>> >> >> I stop it with "killall -9 pgpool" and now wd_lifecheck worked fine<br>
>> on<br>
>> >> >> pgpool-02!<br>
>> >> >> I dont now what was the problem early, but now is working!<br>
>> >> >><br>
>> >> ><br>
>> >> > Maybe you fix some small issue in the config during your testing.<br>
>> >> ><br>
>> >> >> log on pgpool-02:<br>
>> >> >> 2013-08-05 17:52:42 LOG: pid 11524: wd_lifecheck: lifecheck failed<br>
>> 3<br>
>> >> >> times. pgpool 1 (<a href="http://10.0.0.21:5432" target="_blank">10.0.0.21:5432</a> <<a href="http://10.0.0.21:5432" target="_blank">http://10.0.0.21:5432</a>>) seems not<br>
>> to<br>
>> >> be working<br>
>> >> >> 2013-08-05 17:52:42 LOG: pid 11524: wd_escalation: escalated to<br>
>> >> master pgpool<br>
>> >> >> 2013-08-05 17:52:42 LOG: pid 11524: wd_escalation: escalated to<br>
>> >> >> delegate_IP holder<br>
>> >> >> 2013-08-05 17:52:52 LOG: pid 11524: wd_lifecheck: lifecheck failed<br>
>> 3<br>
>> >> >> times. pgpool 1 (<a href="http://10.0.0.21:5432" target="_blank">10.0.0.21:5432</a> <<a href="http://10.0.0.21:5432" target="_blank">http://10.0.0.21:5432</a>>) seems not<br>
>> to<br>
>> >> be working<br>
>> >> >><br>
>> >> >> So rest only one problem, that is how to down delegate_ip from the<br>
>> >> pgpool-01?<br>
>> >> >> Because both servers is responding to delegate_ip.<br>
>> >> ><br>
>> >> > Well, the reason it doesn't get removed on pgpool-01 is because the<br>
>> >> killall -9<br>
>> >> > kills the pgpool processes including the watchdog without any hope of<br>
>> >> them<br>
>> >> > running the ifconfig down command.<br>
>> >> ><br>
>> >> > That said, you just need to run the ifconfig down command on<br>
>> pgpool-01.<br>
>> >> ><br>
>> >> > I'm sure what you're trying to simulate is a crash, but I'm not sure<br>
>> >> killing<br>
>> >> > ALL the pgpool processes with -9 is a good simulation, because more<br>
>> >> likely<br>
>> >> > only one of the backends would crash.<br>
>> >> ><br>
>> >> > Maybe one of the other folks on the list can suggest a better<br>
>> simulation<br>
>> >> for a<br>
>> >> > crashing pgpool service.<br>
>> >><br>
>> >> I have tried with pgpool-II 3.3.0 to test the case. Initially<br>
>> >> "server2" is the watchdog active, and "server1" is the watchdog<br>
>> >> "standby".<br>
>> >><br>
>> >> I killed server2 pgpool-II parent process by kill -9.<br>
>> >><br>
>> >> - server2 releases the VIP. server2 watchdog goes to "down" status.<br>
>> >><br>
>> >> - server1 becomes active and grab the VIP.<br>
>> >><br>
>> >> So my guess is, 3.2's watchdog is not capable to handle the situation.<br>
>> >><br>
>> >> Pgpool-II 3.3's watchdog is much more enhanced than 3.2's. I recommend<br>
>> >> to use 3.3 if you want to seriously use watchdog.<br>
>> >> --<br>
>> >> Tatsuo Ishii<br>
>> >> SRA OSS, Inc. Japan<br>
>> >> English: <a href="http://www.sraoss.co.jp/index_en.php" target="_blank">http://www.sraoss.co.jp/index_en.php</a><br>
>> >> Japanese: <a href="http://www.sraoss.co.jp" target="_blank">http://www.sraoss.co.jp</a><br>
>> >><br>
>><br>
</div></div></blockquote></div><br></div>